November 11, 2008

Socially Conservative Black Voters

In a comment a few days ago, a Reader asked if my observation about Black voters causing Prop. 8 to pass can legitimately be expanded to say that such people are actually socially conservative. The metric, it was suggested, was whether these voters also voted “yes” on Prop. 4, an initiative which failed and would have, if enacted, put restrictions on the ability of a minor to get an abortion without parental consent.

This, it seems to me, was an excellent question and a good measure of it. So I did a little looking around, and I found the Sacramento Bee’s exit poll on Prop. 4. To put the Reader's suggestion to the test, I have culled out not one but two two relevant statistics: race, and prop. 8 voting.

I've done this because the question is whether Black voters are socialy conservative. Prop. 4 and Prop. 8 provide the two easiest measures of social conservatism available on the slate of initiatives that appeared on the California ballot this year. Prop. 8, as you all know, amended the state's Constitution to foreclose the possibility of a same-sex marriage. A "yes" vote is considered socially conservative, and a "no" vote is considered socially liberal. Prop. 4, as you may not know, was an initiative statute which would have required parental notification and a waiting period before a minor obtains an abortion. A "yes" vote is the pro-life, or socially conservative, position; a "no" vote, the opposite.

So I hypothesize that a social conservative would have voted "yes" on both Prop. 4 and Prop. 8. The converse is true also, a social liberal would have voted "no" on both initiatives.

This leaves me with a bit of a taxonomic dilemma, because there are people who voted "yes" on one initiative and "no" on another. Applying the label "moderate" to a pro-parental notification but pro-SSM voter might be technically accurate but it is not particularly informative -- especially because that results in the exact same label being applied to someone whose voting pattern is diametrically opposite that.

Now, to the data. The average “Yes on 4” (that is, pro-parental notification law) voter voted “Yes on 8” 71% of the time, and “No” 29% of the time. The average “No on 4” voter (that is, the ones who did not want to put a parental notification requirement on a minor seeking an abortion) voted almost exactly opposite – these “pro-choice” voters voted “Yes” on 8 23% of the time and “No on 8” 77% of the time.

This is interesting because it shows an almost exact mirror image. If we assume that “Yes on 4” voters have a belief set which is informed by some combination of religion and social conservativism, we would expect them to vote “Yes on 8,” as well. And that is exactly what happened – with the corollary that those whose values on abortion are not informed by religion or social conservatism would be expected to vote against parental notification and in favor of same-sex marriage.

So now, let’s look back at the racial hypothesis. Here, we see that whites voted 45-55 on Prop 4 (“yes” votes come first), blacks voted 51-49, Latinos 53-47, Asians 57-43, and no data is given for other ethnic groups. The overt difference in racial voting supports, but not by a whole lot, the idea that Blacks are socially conservative – assuming that a parental notification law is socially conservative. A very slim majority of Blacks liked the idea. This leaves us with what could a three-dimensional chart, indexing two kinds of voting patterns with race. To go further, I need to make two assumptions and a disclaimer:

Doing the math, the results are as follows (table includes raw vote extrapolation and percentage by racial group):


Racial GroupYes on
both 4 and 8
Yes on 8,
No on 4
No on 8,
Yes on 4
No on
both 4 and 8
White2,314,670
35.77%
856,111
13.23%
759,046
11.73%
2,541,154
39.27%
African-American524,869
51.10%
194,129
18.90%
199,861
6.90%
237,269
23.10%
Latino715,320
38.69%
264,571
14.31%
199,861
10.81%
669,099
36.19%
Asian220,445
35.77%
81,534
13.23%
72,290
11.73%
242,015
39.27%
Other114,721
37.23%
42,431
13.77%
34,728
11.27%
116,262
37.73%

A few notes about how I assembled this data:

First, I assumed that the 71-29 or 23-77 split holds true, on average, across racial lines. Ths is likely not accurate. The question, which I cannot answer, is whether these splits are close enough across racial lines to use those splits as approximations. I do so, but there's your caveat -- my data is not cross-indexed enough, and extrapolating that result from the available information is the point of this exercise.

Second, I assume that every voter actually cast a vote on both initiatives. This is definitely not accurate. The Secretary of State (whose website is finally working again) shows a total of 10,703,198 votes cast on Prop. 4 and 11,084,210 votes cast on Prop. 8 - a difference of more than 380,000 votes. However, I deem the difference of 3.4% to be, while not negligible, not significant enough to render the exercise useless.

Finally, I note that no data was included in Bee's exit poll concerning "other" racial groups (like Indian-Americans, or Native Americans, for instance). So my extrapolation for them is based only on the at-large exit poll numbers.

Simplified to rounded percentages, it looks like this:

Racial Group

Yes on both

Yes on 8, No on 4No on 8, Yes on 4

No on both

White36%13%12%39%
African-American51%19%7%23%
Latino39%14%11%36%
Asian36%13%12%39%
Other37%14%11%38%

Overall, white and Asian voters appear the most socially liberal. The differences between that bloc and combined Latino and "other" voters are very small, nearly insignficant. But standing in great contrast to this are the African-American voters. The proportion of African-American voters in the most conservative category is much higher than in other ethnic groups, and the proprotion of them in the most liberal category is much lower.

One interesting similarity appears, though. Across all ethnic groups, the clumsily-named "moderates," who voted yes on one of the initiatives and no on the other, are all just over a quarter of the total votes cast (about 26%). This is a reflection of the near mirror-image splits on the overall vote. If I could get better data on racial differences between the voting, I would predict more variability in this area of analysis.

But, my Reader's suggestion that African-American voters are particularly intolerant of homosexuality finds some statistical purchase, in that the African-American "split" votes weigh much more heavily in the "No on 4, Yes on 8" category than do the other racial groups.

I conclude from that my Reader and I are both right. As suggested, there does seem to be a measurably more intense strain of opposition to same-sex marriage within African-American voters. And, it is also true that a majority of African-American voters are in the most socially conservative category. Which dovetails back to my original thesis -- these voters were recruited from and motivated by churches and other religiously-affiliated groups and as a result can be expected to be socially conservative, all other things being equal.

This was a very unusual election indeed; it may well be that a significant national political realignment, affecting both parties and all sorts of conventional wisdom, has taken place, or is still taking place.

Now, please bear in mind that I am not, in this post, attempting to blame any group of people for Prop. 8 passing. The blame for that rests with the "No on 8" campaign and not with any ethnic group. But I am trying to understand what is happening politically. What I am observing here is that contrary to the traditional wisdom upon which I was raised and educated, African-American voters (not necessarily all African-Americans, mind you, the voters) have become among the most socially conservative voters out there.

That is not a normative statement. The poll numbers and the election returns are what they are; math is what it is, and I've done the math without consciously looking for a particular result. What I have done here is make a political science observation of moderate sophistication, one that I have endeavored to make in an open-eyed and unprejudiced way. That is all. What we do with this data, whether collectively as a society, as a political party, as a political movement, or as individuals, is a different question.

5 comments:

  1. After I read all your disclaimers at the end, I wasn't sure what you said anymore.

    ReplyDelete
  2. 1. Although I took my best shot, I did have to make some assumptions about voter behavior that are not necessarily accurate.

    2. I am not a racist.

    ReplyDelete
  3. Thank you for your response to my question.

    ReplyDelete
  4. These data and your analyses set up the rationale for new studies to uncover "Why". I suspect the reason is religion. http://pewresearch.org/databank/dailynumber/?NumberID=493

    ReplyDelete
  5. Let's say blacks in CA are homophobes and leave it at that.

    ReplyDelete

Thoughtful, insightful, or informative comments are always welcome. Advertising will be deleted permanently. TL reserves the right to delete comments in his sole discretion (but rarely does so other than for advertising).